© American Association for Cancer Research

Natural history and clinical characteristics of ERBB2 mutant hormone receptor-positive breast cancers: Results from the AACR Project GENIE Registry M. LeNoue-Newton, S.C. Chen, T. Stricker ,D. Hyman, P. Bedard, F. Meric-Bernstam, R. Punglia, D. Schrag, E. Lepisto, F. Andre, L. Smyth, S. Dogan, C. Yu, C. Wathoo, M. Levy, G. Mann, AS. Lalani, F. Ye, C. Micheel and Monica Arnedos

2 www.aacr.org/genie

Disclosures

• Honoraria from Novartis, Astra Zeneca, PUMA, Pfizer and Seattle Genetics

• Research grant from Pfizer and Eli Lilly • Travel expenses from Astra Zeneca, Pfizer and Roche • My institution has received funding for clinical research from

Daichii, Novartis, Roche, Eli Lilly, Pfizer, Merus, MSD, Astra Zeneca

3 www.aacr.org/genie

Background

• Activating ERBB2 mutations have been identified in around 1-2% of breast cancers1,2

• They have been reported to be oncogenic and resistant to some anti-HER2 therapies1 put potentially sensitive to the irreversible TKI neratinib

• Due to the rarity of ERBB2 mutations, evaluation of the natural history of ERBB2 mutant breast cancer requires a large, multi-center series

• The AACR-GENIE consortium database includes information from > 60,000 de-identified genomic records from different types of cancer including nearly 8,700 patients with breast cancer

1. Bose, R. et al. Cancer Discov. 2013; 3: 224–37 2. Ross, J. S. et al. Cancer 2016; 122: 2654–2662

4 www.aacr.org/genie

Methodology

• Multi-center, retrospective, case-controlled study • We interrogated the AACR-GENIE database to

identify HR+/HER2- MBC cases with ERBB2 mutation until end of December 2016

• The objective was to describe the clinicopathological features, response to standard therapies and outcome in the HR+HER2- MBC population harboring an ERBB2 mutation

5 www.aacr.org/genie

Methodology

• Eligibility: • Patients metastatic invasive breast carcinoma • HR-positive, HER2-negative in at least one biopsy sample • Known ERBB2 mutation:

• S310F, S310Y, L755S, L755P, D769H, D769Y, D769N, A775_G776insYVMA, G776delinsVC (G776VinsC), V777L, G778_S779insCPG, P780_Y781insGSP (G778_P780dup), V842I, and L869R

• Matching control cases (2:1) were identified from the database with

known ERBB2-WT and they were matched to the ERBB2mut cases on race, gender, birth year, and age at sequencing at this order

6 www.aacr.org/genie

Study Objectives

• Primary Endpoint: Overall Survival from date of metastatic relapse

• Secondary Endpoints: • Differences clinical and pathological characteristics • OS from diagnostic of primary disease and from date of second line of

treatment metastatic setting • For each line of therapy:

• Duration of therapy • Time to next therapy • Time to progression • Objective Response Rate

7 www.aacr.org/genie

Statistical methods

• For single time-point data, the paired t-test or Wilcoxon signed-rank test or McNemar’s Chi-square

• Between-group comparisons assessed using either analysis of variance

(ANOVA) with adjusted least squares means or Fisher’s exact test • Survival outcome and time-to-event data evaluated by constructing

Kaplan-Meier curves and compared between ERBB2-mut and ERBB2-wt patient groups by log-rank tests

• The Cox proportional hazards model for adjusted tests of significance

and estimates of hazard ratios

8 www.aacr.org/genie

Results

9 www.aacr.org/genie

Sample Characteristics

12 patients had more than one sample sequenced

(6 ERBB2mut, 6 ERBB2wt)

Sequencing in ERBB2mut cases was more frequently conducted on

recurrent/metastatic samples than in ERBB2wt

(70% vs 51%)

10 www.aacr.org/genie

Clinical characteristics at diagnoses between ERBB2mut vs ERBB2wt

No significant differences in major clinicopathological features between

ERBB2mut vs ERBB2wt

11 www.aacr.org/genie

ERBB2mut vs ERBB2wt: • Higher frequency

liver metastases (53% vs 38%, p=0.006) bone metastases (76% vs 64%, p=0.019)

• Lower frequency lung metastases (11% vs 23%, p=0.028)

Overall no differences frequency visceral metastases (64% vs 53%,p=0.37 ERBB2mut vs ERBB2wt, respectively)

Clinical characteristics at relapse between ERBB2mut vs ERBB2wt

12 www.aacr.org/genie

OS analyses

OS from metastatic relapse OS from date of primary diagnoses

Median OS: 55,1mo ERBB2wt

Median OS: 128mo ERBB2mut

Median OS: 134mo ERBB2wt

Median OS: 50,6mo ERBB2mut

13 www.aacr.org/genie

No differences in terms of ORR ascertained by clinical notes between ERBB2mut vs ERBB2wt

ERBB2 mutation and benefit from treatment

Median of lines treatment in the metastatic setting was 5.0 for both ERBB2mut and ERBB2wt (P=0.67)

14 www.aacr.org/genie

ERBB2 mutation and benefit from treatment

• No differences in TTP between ERBB2-mut vs ERBB2wt in first-line treatment (p= 0.8597)

• Similarly in second-line (p = 0.9226)

TTP 1st line of endocrine treatment by ERBB2 mutation status

Median 9.7m ERBB2 wt

Median 8.3m ERBB2mut

• No differences in TTP between ERBB2-mut vs ERBB2wt in first-line endocrine therapy (p= 0.4969)

15 www.aacr.org/genie

Co-Mutations (46 common genes)

ERBB2-mutated ERBB2-wt

No statistically significant differences in most frequent molecular alterations PIK3CA (48% vs. 43%), TP53 (18% vs. 30%) and CDH1 mutations (28% vs. 10%)

N=39

16 www.aacr.org/genie

Co-Molecular aberrations Large Panels139 genes

ERBB2-mutated ERBB2-wt

CDH1 mutation enriched in ERBB2mut (28%) vs ERBB2wt (10%) (p=0.07)

N=24

17 www.aacr.org/genie

ERBB2 mutant population

N=45

ERBB2_variant

L755S 38% (17)

V777L 25% (11)

D769Y/H 15% (7)

S310F/Y 10% (5)

L869R 9% (4)

P780_Y781insGSP 2% (1)

Gao et al. Sci. Signal. 2013 & Cerami et al. Cancer Discov. 2012.

6 ERBB2mut cases were sequenced more than once 3 showed differences: • Two gain ERBB2 mutation at

distant metastases • One lost ERBB2 mutation in later

metastatic setting (untreated with anti-HER2 therapy)

18 www.aacr.org/genie

ERBB2 mutant population

• A total of 19 patients with ERBB2mut were treated with anti-HER2 therapy • Of those, 14 received neratinib, 3 received lapatinib, and 2 received an

undisclosed HER2 TKI • Neratinib has shown efficacy in ERBB2-mutated tumors including breast cancer1

1. Hyman et al, Nature. 2018;554:189-194

No differences in OS between ERBB2mut and ERBB2wt when excluding neratinib-treated patients

Median OS: 55.5mo ERBB2wt

Median OS: 58.9mo ERBB2mut

19 www.aacr.org/genie

Effect neratinib treatment in ERBB2mut patients

• Median duration on neratinib treatment was 148 days and median line treatment administration neratinib in the metastatic setting was 6.0

• Neratinib ORR (CR+PR) = 5.9% with

CBR (CR+PR+SD>24weeks) = 53% • There seems to be a trend towards

improved OS in neratinib treated patients vs no treated although not significant

Median OS 62.8m neratinib treated

Median OS 47.5m neratinib-naïve

20 www.aacr.org/genie

Conclusions (I)

• Limitations of this study: • Retrospective series with low number of cases • Patients with already distant metastases; no information about

negative impact ERBB2mut in DDFS or RFS as other publications1-3 • ERBB2 mutation status categorically classified as present or absent

and type without information allele frequency and multiple platforms

• Strengths: • Largest series so far to describe HR-positive, ERBB2-muttant

population in BC • Data had been compared to matched cases • CLIA-/ISO-certified genomic data

1. Griffith et al. Nature Communications 2018; 9:3476 2. Wang et al, Cancer Science 2017; 108:671 3. Jongen et al, BCRT 2019; 174:55

21 www.aacr.org/genie

Conclusions (II)

• No significant differences in clinicopathological features between ERBB2mut and ERBB2wt tumors except higher rates of bone and liver metastases in ERBB2mut cases and lung metastases in ERBB2wt

• No significant differences observed in OS from diagnostic of distant

metastasis between ERBB2mut and ERBB2wt • ORR and TTP first and second line of treatment did not differ between

ERBB2mut vs ERBB2wt irrespective of the type of therapy • Although some numerical variations, no significant differences in

mutation rates in PIK3CA, TP53, CDH1 or CCND1 Amplification between cases and controls although CDH1 mutation enriched in ERBB2mut (28%) vs ERBB2wt (10%) (p=0.07) and no ESR1mut observed in ERBB2mut

22 www.aacr.org/genie

Conclusions (III)

• Frequency type of ERBB2-mutation according to previously published • The presence of ERBB2 mutation might evolve over time • Subgroup analyses patients treated with neratinib we observed a non-

significant trend on OS for neratinib treatment although study not designed to answer this question (SUMMIT Trial, NCT01953926)

23 www.aacr.org/genie

Acknowledgements

Marta Jimenez Alexandra Jacquet Thomas Filleron

Thomas Bachelot Olivier Tredan

Funda Meric-Bernstam Kenna Shaw

Chetna Wathoo Kristin Hargrave

Fabrice André Semih Dogan

Suzette Delaloge Fadi Sabbagh

Cristophe Massard Ludovic Lacroix

David Hyman Lillian Smyth

Rinaa Sujata Punglia Eva Lepisto

Deborah Schrag

Christine Micheel Michele LeNoue-Newton

Thomas Stricker Mia Levy

Fei Ye Sheau-Chiann Chen

Philippe Bedard Celeste YU

Shawn Sweeny